1 / 17

Hunting for Sharp Thresholds

Hunting for Sharp Thresholds. Ehud Friedgut Hebrew University. Local properties. A graph property will be called local if it is the property of containing a subgraph from a given finite list of finite graphs. (e.g. “Containing a triangle or a cycle of length 17”.). approximable.

zenda
Télécharger la présentation

Hunting for Sharp Thresholds

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Hunting for Sharp Thresholds Ehud Friedgut Hebrew University

  2. Local properties A graph property will be called localif it is the property of containing a subgraph from a given finite list of finite graphs. (e.g. “Containing a triangle or a cycle of length 17”.)

  3. approximable by a local property. Almost- Theorem: If a monotone graph property has a coarse threshold then it is local. Non-

  4. Applications • Connectivity • Perfect matchings in graphs • 3-SAT hypergraphs Assume, by way of contradiction, coarseness.

  5. Generalization to signed hypergraphs Use Bourgain’s Theorem. Or, as verified by Hatami and Molloy: Replace G(n,p) by F(n,p), a random 3-sat formula, M by a formula of fixed size etc.; (The proof of the original criterion for coarseness goes through.)

  6. Initial parameters • It’s easy to see that 1/100n < p < 100/n • M itself must be satisfiable • Assume, for concreteness, that M involves 5 variables x1,x2,x3,x4,x5 and that setting them all to equal “true” satisfies M.

  7. Restrictive sets of variables We will say a quintuple of variables {x1,x2,x3,x4,x5}is restrictive if setting them all to “true” renders F unsatisfiable. Our assumptions imply that at least a (1-α)-proportion of the quintuples are restrictive.

  8. Erdős-Stone-Simonovits The hypergraph of restrictive quintuples is super-saturated: there exists a constant β such that if one chooses 5 triplets they form a complete 5-partite system of restrictive quintuplets with probability at least β. Placing clauses of the form ( x1 V x2 V x3) on all 5 triplets in such a system renders F unsatisfiable!

  9. Punchline Adding 5 clauses to F make it unsatisfiable with probability at least β2{-15}, so adding εn3p clauses does this w.h.p., and not with probability less than 1-2α. Contradiction!

  10. Applications • Connectivity • Perfect matchings in hypergraphs • 3-SAT

  11. Semi-sharp sharp . Rules of thumb: • If it don’t look local - then it ain’t. • No non-convergent oscillations.

  12. A semi-random sample of open problems: • Choosability (list coloring number) • Ramsey properties of • random sets of integers • Vanishing homotopy group • of a random 2-dimensional • simplicial complex.

  13. A more theoretical open problem: • F: Symmetric properties with • a coarse threshold have high • correlation with local properties. • Bourgain: Generalproperties • with a coarse threshold have • positive correlation with local properties. What about the common generalization? Probably true...

  14. Thanks for your attention...

More Related